AMD Radeon RX 560 XT vs AMD Radeon RX 590 GME Comparison

The AMD Radeon RX 560 XT and the AMD Radeon RX 590 GME are two end-of-life Polaris-generation graphics cards that share a common architectural foundation but are positioned very differently in terms of raw performance. The benchmark data presents a clear hierarchy: the RX 590 GME wins both recorded head-to-head tests, yet the RX 560 XT holds a higher overall percentile ranking (79th vs. 77th), a quirk that invites a closer look at how synthetic averages can diverge from direct comparisons.

Where Each One Wins

The use-case split is stark because the RX 590 GME dominates both available benchmark tests. In Geekbench OpenCL, the RX 590 GME scores 36,294 against the RX 560 XT's 31,387, a 13.5% advantage. The gap widens dramatically in Geekbench Vulkan, where the RX 590 GME posts 60,508 versus 36,879 for the RX 560 XT, a 39.1% lead. For any workload that leverages Vulkan — typically modern games and compute APIs — the RX 590 GME is the clear choice. The OpenCL result, while narrower, still favors the RX 590 GME for general-purpose compute tasks.

However, the RX 560 XT is not without a niche. Its average benchmark score of 34,133 is actually higher than the RX 590 GME's 32,601, and its 79th percentile versus the RX 590 GME's 77th percentile suggests that when measured against the entire GPU landscape, the RX 560 XT sits in a slightly better relative position. This could imply that the RX 560 XT has more consistent performance across a broader set of unseen workloads, even if it loses the two tests that were directly compared. For users prioritizing a known quantity in OpenCL and Vulkan, the RX 590 GME wins outright. But for those looking at aggregate standings, the RX 560 XT's average score of 34,133 edges out its rival's 32,601.

Architecture Differences

Both cards are built on GCN 4.0 architecture and use the same 14 nm process node from GlobalFoundries, with an identical transistor count of 5,700 million and a die size of 232 mm². The transistor density is also the same at 24.6M / mm². The fundamental difference lies in the chip configuration: the RX 560 XT uses the Ellesmere chip, while the RX 590 GME uses Polaris 20. This chip difference translates directly into a different number of compute units.

The RX 560 XT has 1,792 shading units, 112 texture mapping units (TMUs), and 32 raster operation units (ROPs). The RX 590 GME bumps shading units to 2,304 and TMUs to 144 while keeping ROPs at 32. This 28.6% increase in shading units and 28.6% increase in TMUs explains most of the performance delta. Clock speeds also differ substantially: the RX 560 XT runs at a base of 1074 MHz and a boost of 1226 MHz, whereas the RX 590 GME operates at a base of 1257 MHz and a boost of 1420 MHz. The RX 590 GME's higher clocks, combined with more compute units, produce a theoretical FP32 throughput of 6.543 TFLOPS versus 4.394 TFLOPS for the RX 560 XT. Both cards support FP16 at a 1:1 ratio.

Memory configurations also diverge. The RX 560 XT comes with 4 GB of GDDR5 on a 256-bit bus, running at 1750 MHz (7 Gbps effective) for 224.0 GB/s of bandwidth. The RX 590 GME doubles the capacity to 8 GB of GDDR5 on the same 256-bit bus, but runs the memory at 2000 MHz (8 Gbps effective), yielding 256.0 GB/s. The 8 GB capacity is a significant advantage for modern games that exceed 4 GB VRAM usage, and the 14.3% higher bandwidth helps texture-heavy scenes.

FAQ

Q: Which card has a higher average benchmark score?

A: The AMD Radeon RX 560 XT has an average benchmark score of 34,133, which is higher than the AMD Radeon RX 590 GME's average of 32,601.

Q: Does the RX 590 GME win every head-to-head benchmark?

A: Yes, the RX 590 GME wins both recorded tests. It leads by 13.5% in Geekbench OpenCL and by 39.1% in Geekbench Vulkan.

Q: What is the VRAM capacity difference?

A: The RX 560 XT has 4 GB of GDDR5 memory, while the RX 590 GME has 8 GB of GDDR5 memory. Both use a 256-bit bus.

Q: Are both cards based on the same architecture?

A: Yes, both use GCN 4.0 architecture, are built on a 14 nm process, and share the same transistor count of 5,700 million and die size of 232 mm².

Q: Which card has a higher boost clock?

A: The RX 590 GME has a boost clock of 1420 MHz, compared to the RX 560 XT's boost clock of 1226 MHz.

Q: How do their percentile rankings compare?

A: The RX 560 XT is in the 79th percentile of all GPUs, while the RX 590 GME is in the 77th percentile.

Specification Differences

The key specification differences are concentrated in compute resources, clock speeds, memory, and power delivery. The RX 590 GME has 2,304 shading units versus 1,792 for the RX 560 XT, and 144 TMUs versus 112. Both have 32 ROPs. Base clocks are 1257 MHz for the RX 590 GME and 1074 MHz for the RX 560 XT; boost clocks are 1420 MHz and 1226 MHz, respectively.

Memory size is 8 GB on the RX 590 GME and 4 GB on the RX 560 XT. Memory clocks are 2000 MHz (8 Gbps effective) versus 1750 MHz (7 Gbps effective), producing bandwidth of 256.0 GB/s versus 224.0 GB/s. The pixel rate is 45.44 GPixel/s for the RX 590 GME and 39.23 GPixel/s for the RX 560 XT. Texture rate is 204.5 GTexel/s versus 137.3 GTexel/s. FP32 compute is 6.543 TFLOPS versus 4.394 TFLOPS.

Power draw and connectors also differ. The RX 590 GME has a TDP of 175 W and requires a single 8-pin power connector, while the RX 560 XT has a TDP of 150 W and uses a single 6-pin connector. Both suggest a 450 W PSU. Physical dimensions are identical at 241 mm (9.5 inches) in length, both are dual-slot, and both use PCIe 3.0 x16. Display outputs are the same: 1x HDMI 2.0b and 3x DisplayPort 1.4a. The RX 560 XT was released on March 12, 2019, while the RX 590 GME followed on March 8, 2020.

Head-to-Head Benchmarks

The most decisive result is in Geekbench Vulkan, where the RX 590 GME scores 60,508 against the RX 560 XT's 36,879. The delta of -39.1% (from the RX 560 XT's perspective) means the RX 590 GME is roughly 64% faster in this test. This is a massive margin that points to the RX 590 GME's superior shading unit count and higher memory bandwidth being particularly impactful in Vulkan's low-overhead path.

In Geekbench OpenCL, the gap narrows considerably but remains one-sided. The RX 590 GME scores 36,294, and the RX 560 XT scores 31,387, a 13.5% difference. This suggests that OpenCL workloads are less sensitive to the RX 590 GME's extra compute units and more dependent on other factors like clock speed. Still, the RX 590 GME wins.

What is curious is the inversion in average scores. The RX 560 XT's average across all benchmarks is 34,133, which is 1,532 points higher than the RX 590 GME's 32,601. This happens because the RX 590 GME's average includes a single 3DMark Steel Nomad DX12 result of just 1,001, which pulls its average down significantly. The RX 560 XT has no such low-scoring entry. This teaches an important lesson: head-to-head wins in specific APIs do not always translate to a higher overall average when other tests are included.

The Verdict

The data is unambiguous for anyone prioritizing the two benchmark tests that were directly compared. The AMD Radeon RX 590 GME is the superior card for Geekbench OpenCL and Geekbench Vulkan workloads, with leads of 13.5% and 39.1% respectively. Its 8 GB VRAM capacity, higher 1420 MHz boost clock, and 256.0 GB/s bandwidth make it the stronger choice for modern, memory-hungry applications.

However, the AMD Radeon RX 560 XT is not without merit. Its higher average benchmark score of 34,133 and superior 79th percentile ranking suggest it may perform more consistently across the wider universe of GPU tests, even if it loses the two that were measured head-to-head. For users whose workloads are not dominated by Vulkan or OpenCL, the RX 560 XT's lower 150 W TDP and single 6-pin connector make it an easier fit for systems with modest power supplies, even though both cards recommend a 450 W PSU.

The verdict comes down to workload specificity. If the software in question uses Vulkan, the RX 590 GME is the only rational choice — a 39.1% performance advantage is too large to ignore. If the workload is unknown or varied, the RX 560 XT's higher aggregate standing and lower power draw (150 W vs. 175 W) make it a defensible option. The data does not support a universal winner; it supports a decisive winner for specific APIs and a marginal winner for overall consistency.